Shelter Disinflation Meets Goods Reflation in March, and Immaculate Disinflation

Consumer price inflation (CPI) fell below consensus expectations in March, with headline CPI advancing at a softer pace though it matched our forecast. On the other hand, prices in the core segment lost momentum. While the gain in the core index met our and the consensus forecast, it also registered its softest increase in four months (0.385% m/m unrounded). Importantly, March's weakish core CPI reading reflected the m/m slowing in the services segment, which saw a surprising deceleration in shelter prices.

Core price inflation is still running at a hot 5.1% three-month annualised pace, barely unchanged vs February. In our view, the path for core inflation over the next few months will be determined by the tug-of-war between rising momentum in goods prices (used vehicles) and slowing services inflation (softer shelter). Fed officials, particularly Chair Powell, have continued to flag core services ex-shelter inflation as the main indicator to track in order to gauge whether underlying inflation has begun to turn the corner. The forthcoming March PCE report will tell us more about these dynamics.

We continue to anticipate core inflation will stay largely around the 0.4% m/m range in April and May. In our view, the short-term outlook is likely to give pause to Fed officials and leaves 25bp rate increases on the table for both the May and the June FOMC meetings despite recent financial stress caused by the banking crisis.

In separate analysis, we use a time-varying Phillips curve to interpret recent core Personal Consumption Expenditure (PCE) inflation dynamics and build a projected path for 2023 and 2024. More specifically, we look at the relationship between inflation minus its trend ("inflation gap") vs the unemployment gap based on an estimate of NAIRU (the unemployment rate at which inflation no longer accelerates).

We start with spelling out our assumptions regarding the future unemployment rate path: Given a persistently constrained labor supply with a lower participation rate vs pre-Covid, the level of the unemployment rate will be restrained. Given this, and assuming that the economy slides into a recession similar in severity to the 1990-1991 and 2001 recessions, we expect the unemployment rate to accelerate from 3.5% currently to 4.0% by year-end, 5.5% by Q4 2024, and peaking at 5.6% in Q1 2025.

The time-varying Phillips curve framework provides some interesting facts with regards to changing inflation dynamics over the recent period. Firstly, the long-term trend inflation has drifted up since 2020 to above-target levels for core PCE inflation. In conjunction with higher trend inflation, we've also seen an increase in the aggregate sensitivity of core PCE inflation to labor market tightness. When benchmarked against the Fed's 2% inflation target, current inflation dynamics seem to suggest that the likelihood of a return of long-term inflation trend towards 2% is relatively low. This puts a floor under how much disinflation we can expect to see, even with a recession hitting near term.

Aggregating our results, we find that continued labor market strength will keep core PCE inflation sticky in 2023, ending the year just below 4%, and the recessionary impulse will bring it just below 3% by end-2024. However, this assumes a gradual return of long-term inflation trends in line with the 2% target by the end of 2025. If these trends remain above target, it is unlikely we will see core PCE inflation dropping below 3%. Irrespective of this assumption about trend inflation, core PCE inflation remains sticky this year on the back of labor market strength, which likely means that the Fed has more hikes to come and will be very wary to cut rates throughout the remainder of 2023.